SCRS/2019/188 Collect. Vol. Sci. Pap. ICCAT, 76(10): 365-373 (2020) EVALUATION OF POST-RELEASE MORTALITY FOR PORBEAGLE AND SHORTFIN MAKO SHARKS FROM THE CANADIAN PELAGIC LONGLINE FISHERY H. Bowlby1, W. Joyce1, H. Benoit2 and J. Sulikowski3 SUMMARY The majority of shark catches from the Canadian pelagic longline fleet are discarded alive at sea, making post-release mortality (PRM) estimates critical to understanding total fishing mortality. We found little evidence that at-vessel mortality has changed from 2001-2018, suggesting that capture characteristics are similar and previous satellite tagging can be combined with more recent tagging to describe PRM. Estimated rates were 14% for porbeagle (6% for healthy and 40% for injured) and 28% for shortfin mako (27% for healthy and 33% for injured), which is approximately ½ of the previous estimate for porbeagle and essentially the same for shortfin mako. We propose that this difference for porbeagle is related to handling characteristics during tagging, which switched from bringing animals on board to tagging in the water. This conclusion is supported by an analysis of recovery times for surviving animals where median recovery time was 1 day (shortfin mako) or 1.5 days (porbeagle) longer when the shark was tagged onboard as compared to in the water. RÉSUMÉ La majorité des prises de requins de la flottille canadienne de palangriers pélagiques sont rejetées vivantes en mer, ce qui rend les estimations de mortalité après remise à l'eau (PMR) essentielles pour comprendre la mortalité par pêche totale. Nous n'avons trouvé que peu d'éléments probants indiquant que la mortalité à bord des navires a changé entre 2001 et 2018, ce qui donne à penser que les caractéristiques de capture sont semblables et que le marquage par satellite antérieur peut être combiné avec un marquage plus récent pour décrire la PMR. Les taux estimés étaient de 14% pour le requin-taupe commun (6% pour les spécimens en bonne santé et 40% pour les spécimens blessés) et de 28% pour le requin-taupe bleu (pour les spécimens en bonne santé et 33% pour les spécimens blessés), soit environ ½ des estimations précédentes pour le requin-taupe commun et essentiellement les mêmes pour le requin-taupe bleu. Nous supposons que cette différence pour le requin-taupe commun est liée aux caractéristiques de manipulation pendant le marquage, qui est passé du hissage des animaux à bord de l’embarcation à leur marquage dans l'eau. Cette conclusion est étayée par une analyse des temps de récupération des animaux survivants pour lesquels le temps de récupération médian était d'un jour (requin-taupe bleu) ou d'un jour et demi (requin-taupe commun) de plus lorsque le requin était marqué à bord que lorsqu’il était marqué dans l'eau. RESUMEN La mayoría de las capturas de tiburones de la flota palangrera pelágica canadiense son descartadas vivas en el mar, lo que hace que las estimaciones de mortalidad posterior a la liberación (PMR) sean críticas para comprender la mortalidad total por pesca. Encontramos poca evidencia de que la mortalidad en los buques haya cambiado entre 2001 y 2018, lo que sugiere que las características de captura son similares y que el marcado vía satélite previo puede combinarse con un marcado más reciente para describir la PMR. Las tasas estimadas fueron del 14 % para el marrajo sardinero (6 % para los sanos y 40 % para los heridos) y del 28 % para el marrajo dientuso (27 % para los sanos y 33 % para los heridos), que es aproximadamente la mitad de la estimación anterior para el marrajo sardinero y esencialmente 1 Fisheries & Oceans Canada, Bedford Institute of Oceanography, Dartmouth, Nova Scotia, CANADA. E-mail address of lead author: [email protected]. 2 Fisheries & Oceans Canada, Institut Maurice Lamontange, Mont Joli, Quebec, CANADA. 3 Department of Marine Sciences, University of New England, Biddeford, Maine, USA 365 la misma para el marrajo dientuso. Proponemos que esta diferencia para el marrajo sardinero está relacionada con las características de manipulación durante el marcado, que pasaron de llevar animales a bordo a marcar en el agua. Esta conclusión se apoya en un análisis de los tiempos de recuperación de los animales supervivientes, donde el tiempo medio de recuperación fue de un día (marrajo dientuso) o 1,5 días (marrajo sardinero) y más largo cuando el tiburón fue marcado a bordo en comparación con el marcado en el agua. KEYWORDS Fishing mortality, live discards, tagging effects, post-release mortality, pelagic sharks 1. Introduction In Canada, shortfin mako and porbeagle sharks are two species where license conditions in the pelagic longline fleet stipulate the mandatory release of all live bycatch, although dead animals may be landed. As live bycatch represents the majority of catches, post-release mortality (PRM) could represent the majority of total fishing mortality by Canadian fleets, as opposed to landings and dead discards (Campana et al. 2016). Given that both species are thought to be at low abundance based on their most recent assessments (ICCAT 2009, ICCAT 2018), there is an immediate need for accurate PRM estimates to account for mortality from discard events when calculating total annual removals. Current estimates for post-release mortality for shortfin mako and porbeagle from pelagic longline fleets are limited. A recent meta-analysis applied a random effects model to the PRM results from 4 studies on shortfin mako and estimated PRM to be 25% (95% CI = 14,42%; Musyl & Gillman 2019). There was no similar estimate for porbeagle as multiple studies have yet to be completed. Previous Canadian estimates of PRM were 27% (10% for healthy sharks, 75% for injured) for all captured porbeagle and 31% (30% for healthy sharks, 33% for injured) for all captured shortfin mako (Campana et al. 2016). Condition was accounted for because it has been found to be a good predictor of subsequent mortality for several commercially fished species (e.g. Benoit et al. 2010, Skomal 2007, Ellis et al. 2017). However, sample sizes for injured animals were extremely low (n = 4 and n = 3, respectively) and estimated rates may not be representative. Several tagging programs for pelagic sharks, including Canada’s, have shifted towards tagging in the water as opposed to bringing animals on board in an effort to minimize handling effects. This affords a unique opportunity to assess how handling influences recovery time from datasets where the method of capture remains consistent. Longer recovery times are associated with processes that cause greater physical and physiological trauma to released animals (Ellis et al. 2017). Stresses associated with handling would be low relative to those associated with the capture process itself and are unlikely to be a large component of PRM (Campana et al. 2009, Musyl & Gilman 2019). However, handling effects are rarely evaluated for satellite tag studies yet have important implications for study design and conclusions on how the capture process affects shark bycatch. 2. Methods 2.1. Tagging Details on the tag types and tagging methodology used previously to evaluate PRM for shortfin mako and porbeagle from the Canadian pelagic longline fishery can be found in Campana et al. (2016). We will refer to that tagging as the early period and our tagging as the recent period for the remainder of this document. We used essentially the same methodology as in Campana et al. (2016), albeit with a different type of satellite tag. In brief, shortfin mako and porbeagle sharks were captured from pelagic longline sets and tagged by fisheries observers or science personnel with short-term archival survival tags (Lotek PSATLIFE). The majority of tagging took place during regular commercial fishing for Swordfish in 2017-2018, with a limited number of porbeagle (n = 11) tagged during a pelagic shark survey in 2017. In all cases, captures occurred from pelagic longline gear fished for a minimum soak time of 6 hours. The fleet switched over to using circle hooks in the early 2000s and all captures in 2017-2018 were on circle hooks. For animals that were tagged in the water, the line was cut as close to the hook as possible for release. For animals that were brought onboard, measurements and tagging took 1-2 minutes while the animal’s gills were being irrigated with seawater. In this instance, hooks were removed before release. The 366 tags were attached to the sharks by angling the dart anchor into the pterygiophores in the dorsal musculature, immediately beside the posterior end of the first dorsal fin (Campana et al. 2016, Musyl et al. 2011). Individuals of both species were chosen opportunistically for tagging, but efforts were made to include a range of sizes and both sexes for each species. PSATLIFE tags record pressure (0.05% resolution), temperature (± 0.2 °C) and light intensity every 5 minutes for the 28-day deployment. Mortalities are inferred from continual records at a constant depth for multiple days (indicative of a dead animal on the bottom) or pop-ups following progressively increasing depth records up to the tag crush depth of 2000 m (Musyl et al. 2011). For these analyses, data collected for the previous study (2001- 2013; Campana et al. 2016) were truncated to 28 days to be comparable. As such, these results only apply to mortality events that occur immediately following capture/handling and would not represent delayed mortality resulting from internal damage causing cessation of feeding. The tagging data gave survival in continuous time (days until death) with right-censored observations from individuals that lived until the end of the observation period (28 days). The observations are censored because the ultimate time of death of the individual is unknown, yet they are known to be alive until the end of the observation period.